Part II - Google PlayStore App Data

by Moyinoluwa Busola Agbaje

Investigation Overview

• 'Everyone' contents is the most (highly) rated content in Content Rating.
• 'Free Installs' are the most preferred compared to Paid installs.
• Number of Free compared to Paid Installs.
• Top ten (10) Categories.
• Top ten (10) Genres.
• Category Concentration.
• Paid Apps (Earnings per Category).
• Relationship between App, Category and Type.

Dataset Overview

The dataset used for this exploration was web scrapped by from Google PlayStore analytics by Lavanya Gupta in 2018. This was provided by companies like Annie or Sensor Tower, which helps to drive development and App marketing strategies to succeed for many companies.

# remove Apps that are greater than > 250

df_gapps_clean = df_gapps_clean[df_gapps_clean['Price'] < 250 ]
df_gapps_clean.sort_values('Price', ascending=False).head()

	App	Category	Rating	Reviews	Size_MBs	Installs	Type	Price	Content_Rating	Genres
2281	Vargo Anesthesia Mega App	MEDICAL	4.60	92	32.00	1000	Paid	79.99	Everyone	Medical
1407	LTC AS Legal	MEDICAL	4.00	6	1.30	100	Paid	39.99	Everyone	Medical
2629	I am Rich Person	LIFESTYLE	4.20	134	1.80	1000	Paid	37.99	Everyone	Lifestyle
2481	A Manual of Acupuncture	MEDICAL	3.50	214	68.00	1000	Paid	33.99	Everyone	Medical
2463	PTA Content Master	MEDICAL	4.20	64	41.00	1000	Paid	29.99	Everyone	Medical

Visualization 1

• 'Everyone' contents is the most (highly) rated content in Content Rating.
• 'Free Installs' are the most preferred compared to Paid installs.
• Top ten (10) Categories.
• Top ten (10) Genres.

content_ratings = df_gapps_clean['Content_Rating'].value_counts()

# plot a pie chart
plt.figure(figsize=(10, 8))
figr = px.pie(labels=content_ratings.index,
             values=content_ratings.values,
             title="Content Rating",
            names=content_ratings.index);

# update traces
figr.update_traces(textposition="outside", textinfo='percent + label');

figr.show()

<Figure size 720x576 with 0 Axes>

paid_vs_free_apps = df_gapps_clean['Type'].value_counts()

# plotting a donut chart
plt.figure(figsize=(10, 8))
fig = px.pie(labels=paid_vs_free_apps.index,
             values=paid_vs_free_apps.values,
             hole=0.6,
             title="Paid Vs Free Apps",
            names=paid_vs_free_apps.index);

# update traces
fig.update_traces(textposition="outside", textinfo="percent + label");

fig.show()

<Figure size 720x576 with 0 Axes>

top_10_categories = df_gapps_clean['Category'].value_counts()

# plot bar chart
plt.figure(figsize=(12, 10))
plt.xticks(rotation=30)
plt.xlabel("Top 10 Categories")
plt.ylabel("Count")
plt.title("Top 10 Categories for Apps")
fig = plt.bar(x=top_10_categories.index[:10], height=top_10_categories.values[:10]);

plt.show()

top_10_genres = df_gapps_clean['Genres'].value_counts()

# plot horizontal bar chart

plt.figure(figsize=(10, 8))
plt.yticks(rotation=30)
plt.xlabel("Count")
plt.ylabel("Genres")
plt.title("Top 10 Genres")

plt.barh(top_10_genres[:10].index, top_10_genres[:10].values);

plt.show()

Visualization 2

• Number of Free Apps compared to Paid Apps Installs.
• Category Concentration.
• Paid Apps (Earnings per Category).

# I will have to create a dataframe to show this relationship, then groupby Category and sum the number of installs

install_category = df_gapps_clean.groupby("Category").agg({"Installs": pd.Series.sum})

# groupby Category and count the number of apps

numb_category = df_gapps_clean.groupby("Category").agg({"App": pd.Series.count})

# the new dataframe created

df_merger = pd.merge(numb_category, install_category, on="Category", how="inner")
df_merger.head()

	App	Installs
Category
ART_AND_DESIGN	61	114233100
AUTO_AND_VEHICLES	73	53129800
BEAUTY	42	26916200
BOOKS_AND_REFERENCE	169	1665791655
BUSINESS	262	692018120

#  plot scatter plot

fig = px.scatter(df_merger, # data 
                x='App', # names
                y='Installs', 
                size="App",
                hover_name=df_merger.index,
                color="Installs", 
                title="Category Correlation")

fig.update_layout(xaxis_title="Number Of Apps",
                 yaxis_title="Installs",
                 yaxis=dict(type="log"))

fig.show()

# plot box plot

fig = px.box(df_gapps_clean, 
             x="Type", 
            y= "Installs", 
            color = "Type", 
            notched=True, 
            title="How many downloads are been Paid for compared to Free installs?", 
            )
fig.update_layout(xaxis_title="Type",
                 yaxis_title="Installs", 
                 yaxis=dict(type=("log")))
fig.show()

# plotting

paid_gapps = df_gapps_clean[df_gapps_clean["Type"]== "Paid"]
    
box = px.box(paid_gapps, 
             x="Category", 
            y= "Revenue_Estimate",  
            title="How much can be earned on Paid Apps per Category?"
            )
box.update_layout(xaxis_title="Category", 
                 yaxis_title="Paid App BallPark Revenue", 
                  xaxis= {'categoryorder': 'min ascending'},
                  yaxis= dict(type=("log")))


box.show()

Visualization 3

• Relationship between App, Category and Type.

# free vs paid apps

free_vs_paid = df_gapps_clean.groupby(["Category", "Type"], as_index=False).agg({"App": pd.Series.count})

# plot a clustured bar chart

plt.figure(figsize=(16, 10))
plt.xlabel("Category")
plt.ylabel("App")
plt.title("Relationship between App, Category and Type.")
plt.xticks(rotation=90)
sb.barplot(data=free_vs_paid, x="Category", y="App", hue="Type");

plt.show()